Scalable solutions for IP rigidity

Abstract

An overlay network of connection ID (called CID) processing nodes, called C-nodes, is deployed to solve joint IP mobility, NAT traversal, and path identity problems, in a unified framework. The CIDs can either be globally unique or unique at local C-nodes. CIDs can be modified along a path, and each connection is enabled to send packets in multiple paths via multiple IP networks. It is also possible to direct a path to traverse through a particular node. Double-sided/single-sided mobility, with double-sided/single-sided NAT traversal problems are solved using C-nodes deployed as terminal software, gateways, or relays.

1. A method to reduce IP (Internet protocol) rigidity in the problems of mobility, NAT (network address translation) traversal and explicit path identity, comprising:
 setting up a plurality of overlay nodes, called C-nodes; 
 a C-node is software inserted into a host or a terminal; 
 a C-node is enabled in three alternate forms: hooking, layer 2.5 module, or router; wherein at a C-node, a connection ID (called CID) may be inserted into or removed from a packet; the CIDs inserted into the packets of a single IP connection are being the same; the CIDs are locally unique in that different IP connections crossing at the same C-node is uniquely differentiated by the CIDs; the CID may contain the MAC (media access control) address of a host or terminal; 
 wherein a soft switch called a C-switch is maintained at each C-node to perform a packet operation (PO) which is any combination of the procedures: (1) receiving an inbound packet, (2) generating an outbound packet, (3) dropping (deleting) an inbound packet, (4) replicating an inbound packet into multiple packets, or (5) forwarding an outbound packet; the PO may be dependent on the CID inserted in the inbound packet; 
 wherein the C-switch at each C-node also performs a CID-operation (CO) which is any combination of the procedures performed on an inbound packet: (1) inserting a CID into the packet, (2) removing a CID from the packet, (3) modifying the packet header, or (4) modifying a CID in the packet; 
 wherein at a C-node, any combination of POs and COs is performed on a packet arriving at the C-node only if the destination IP address in the packet is the same as the IP address of the host or terminal in which the C-node resides; 
 wherein only a select subset of C-nodes are allowed to perform COs or POs on packets arriving at the select C-nodes. 
 
     
     
       2. The method of  claim 1 , wherein a plurality of special C-nodes (called gateway C-nodes) is set up; each gateway C-node resides at a host or terminal with a public IP address; such a host or terminal may be the ingress or the egress point of an autonomous system or an intermediate system; at such a location, a NAT may be present. 
     
     
       3. The method of  claim 2 , wherein a host or terminal without an inserted gateway C-node may be associated with a specific gateway C-node, which is called the associated gateway C-node of the host or terminal. 
     
     
       4. The method of  claim 3 , wherein a host or terminal behind a NAT sends its packets to its associated gateway C-node from behind the NAT, in order to traverse the NAT. 
     
     
       5. The method of  claim 4 , a gateway C-node associated with a host or terminal may be re-associated through a control-plane action between the C-nodes and hosts or terminals. 
     
     
       6. A computer-readable non-transitory medium with a computer program for performing the methods as described in any one of  claims 1  to  5 .